Energy efficient scheduling for deadline-constrained applications in edge computing systems

• Main Author: Wang, Qianteng
• Other Authors: Arvind Easwaran
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; Edge computing; Internet of Things; Scheduling
• Online Access: https://hdl.handle.net/10356/175127
• Institution: Nanyang Technological University

Edge computing is a rapidly advancing computing paradigm that brings computation closer to the location where it is needed, thereby enhancing response time and reducing bandwidth. This approach is particularly beneficial for tasks with stringent deadlines. Exploiting these advantages, our research endeavors to tackle the optimization of end device energy consumption in a multi-layered edge computing framework. Our focus is on the concurrent optimization of service placement, offloading scheduling, and processing scheduling while considering the constraints of limited resources and strict deadlines. This optimization challenge is formulated as an Integer Non-Linear Programming (INLP) problem. To address this problem, we propose a novel local search-based algorithm named Heuristic Based Local Search (HBLS), which decomposes the problem into two subproblems and delivers efficient polynomial-time solutions. Our simulation results reveal that HBLS outperforms the traditional First-Come-First-Serve (FCFS) and Earliest-Deadline-First (EDF) strategies by 25.7% and 29.6% in energy savings, respectively. These results highlight the effectiveness and potential of our approach in enhancing energy efficiency for all users in edge computing environments.